This online eBook provides insight and advice on how to build an effective disaster recovery strategy in the evolving world of virtual infrastructures,...
Off the rack CRM doesn't fit every business. Are you better off with a customized solution that addresses your unique business challenges? This new white...
I think I'll call this kernel the "sticky turtle", in honor of
that historic "greased weasel" kernel, and as a comment on how
sadly dependent I've become on the daily BK snapshots. It's been
too long since 2.5.70.
I'll do better, I promise. While most developers happily use
either the daily snapshots or the BK tree itself, not everybody
does, and a lot of users want "real releases".
There's nothing hugely interesting here, but Al Viro ha sbeen
cleaning up the tty layer, and Stephen Hemminger has been fixing up
some network device alloc/free issues with the help of various
people.
And obviously there are the normal usb/pcmcia/sound/architecture
updates. With driver models and networking thrown in as a
bonus.
Linus
---
Summary of changes from v2.5.70 to v2.5.71
Adrian Bunk:
kbuild: [PATCH] document modules_install in "make help"
[wan lmc] remove 2.0.x-era code
SECURITY_ROOTPLUG must depend on USB
Alan Cox:
[netdrvr tlan] fix 64-bit issues
Alan Iwi:
Missing magic number
Alan Stern:
USB: fix address assignment after device reset
USB: Don't allocate transfer buffers on the stack in hub.c
USB: Rename static functions in hub.c and increase
timeouts
USB: Make hub.c DMA-aware
Alexander Viro:
callout removal: ircomm_tty
tty_driver refcounting
Fix sound lockup - missing chardev init
procfs bug exposed by cdev changes
Fix disk partitioning with multiple IDE disks
>minor_shift removal
[NET]: Eliminate {init,register,unregister}_fcdev
[NET]: Eliminate init_hippi_dev and {un,}register_hipdev
[NET]: Convert most tokenring drivers away from
{init,register,unregister}_trdev, only ibmtr remains
[NET]: Eliminate init_fddidev
[NET]: Move sk98lin driver away from init_etherdev()
[NET]: Move 3c509 driver away from init_etherdev()
[NET]: Move sunqe.c driver away from init_etherdev()
[NET]: Move bmac.c away from init_etherdev()
[NET]: Convert USB drivers away from init_etherdev()
Alexey Kuznetsov:
tcp_output.c, tcp.c, tcp.h
Andi Kleen:
Work around gcc 3.3 bug on amd64 in binfmt_elf.c
Make spinlock debugging compile on x86-64
x86-64 merge
Andreas Schultz:
I2C: fix unsafe usage of list_for_each in i2c-core
Andrew Morton:
truncate and timestamps
fix typo in coda
Fix suspend with pccardd running
fix oops on resume from apm bios initiated suspend
export mmu_cr4_features to modules
[VISWS] irqreturn_t conversion
Fix CONFIG_PROCFS=n
zoran user-pointer fix
irq balance logic fix
kill lock_kernel() in inode_setattr()
i2o memleak comment
write_one_page() fixlets
speed up the unlink speedup
Remove unneeded fcntl check
unregister_netdev cleanups
support 64 bit pci_alloc_consistent
svcsock use-after-free fix
Fix writev when a segment generates EFAULT
Fixes trivial error in
[NET]: Convert rtnl_lock/register_netdevice/rtnl_unlock to
register_netdev
fix generic_file_write()
magazine layer for slab
Additional fields in slabinfo
initialise vma->vm_next in do_mmap_pgoff()
More irq balance fixes
dirty_writeback_centisecs fixes
Copy nanosecond stat values for i386
Some subarch warning fixes
MTD build fix
improved core support for time-interpolation
ext3: fix for blocksize < PAGE_CACHE_SIZE
/proc/kcore fixes
remove 16-bit pid assumption from ipc/sem.c
fix broken networking
[NET]: Fix broken cleanups in net/core/iovec.c
aio: small cleanups
clean up timer interpolation code
radio-cadet.c: remove unnecessary copy_to_user()
cmpci: fix improper access to userspace
zr36120: fix improper access to userspace
remove unsafe BUG_ON()
pnpbios dereferencing user pointer
fix bw-qcam.c bad copy_to_user
Graceful failure in devfs_remove()
Fix generic_file_write() again
reiserfs option parser fix and ability to pass
reiserfs support for blocksizes other than 4096 bytes
hugetlbfs: mount options and permissions
DEADLINE: hash removal fix
ext3: fix deadlock in journal_create()
kmalloc_percpu: interface change
per-cpu support inside modules (minimal)
IRQs: handle bad return values from handlers
IRQs: fix up irq_desc initialisation for non-ia32
force_successful_syscall_return()
fix wobbly /proc/stat:btime
Console blanking fix
Console privacy for braille users
Fix tty devfs mess
misc fixes
cs423x fixes
remove get_current_user()
remove triggerable BUG() from de_thread
Don't let processes be scheduled on CPU-less nodes (1-3)
DAC960 fix for fibre channel transfer rate
/proc/sys/vm/min_free_kbytes
loop: remove the balance_dirty_pages() call
Fix build for CONFIG_KALLSYMS=n
ppc64: fixup for family/sk_family rename
Fix the build with !CONFIG_PROC_FS
common 32-bit ioctl code
ioctl32 cleanup: sparc64
x86_64: use common ioctl code
remove_proc_entry() fix
JFFS_PROC_FS must depend on JFFS_FS
fix apic handling for NUMA-Q
cleanup conditionals in summit subarch
provide bus to node mapping for Summit
rocket.c: devfs fix
add bootmem failure warning
eventpoll: fix possible use-after-free
Remove DRM ioctls for common compat ioctl code
fix possible busywait in rtc_read()
fix discontig with 0-sized nodes
fix TARGET_CPUS inconsistency
update MAINTAINERS for Compaq drivers
optimize fixed-sized kmalloc calls
fix scheduler bug not passing idle
fix numa meminfo
fix oops in driver/serial/core.c
fix hangs with nfs to localhost
devfs_mk_dir() fix
cpqarray.c: fix stack usage
dirty bit clearing on s390
drivers/char/mem.c cleanup
irq_cpustat cleanup
MAINTAINERS: Compaq->HP
loop: file use highmem
loop: make bio_copy private to loop
loop: loop bio renaming
loop: copy bio not data
loop: remove an IV
loop: remove LO_FLAGS_BH_REMAP
loop: remove blk_queue_bounce
loop: copy_bio use highmem
loop: don't lose PF_MEMDIE
tmpfs: shmem_file_write EFAULT
tmpfs: swapoff-truncate race
tmpfs: misc fixes
cleanup seqfile usage in resource.c
x25 facilities parsing fix
eicon usercopy fix
intermezzo symlink fix
mdc800 usercopy fix
mpu401 usercopy fix
emu10k1 memleak fix
rio memleak fix
fix resource leak in i810 driver
tioclinux() numbers in <linux/tiocl.h>
fix writeback for dirty ramdisk blockdev inodes
Andries E. Brouwer:
isa_writeb args interchanged
scsi: ten -> use_10_for_rw / use_10_for_ms
Andy Grover:
ACPI: fix extra semicolon (Pavel Machek)
ACPI: Trivial name init patch (Bjorn Helgaas)
ACPI: Disable ACPI sleep on SMP systems (Pavel Machek)
ACPI: Add access checking (Andi Kleen)
ACPI: Add ACPI PCI Subdriver. This enables acpiphp to work
again. (Matthew Wilcox)
ACPI: Implement PCI Domain support (Matthew Wilcox)
Anton Blanchard:
ppc64: comment fix from Milton Miller
ppc64: Fix problem creating zImage/zImage.initrd multiples
times from Bryan Logan
ppc64: clean up SLB reload code and remove some unnecessary
isyncs
ppc64: fix compile warnings
ppc64: fix misreporting of unhandled IRQs in xics IPI
ppc64: cleanup some hardcoded constants
ppc64: replace MAX_PACAS with NR_CPUS
ppc64: Fix a bad shift against PCI_BASE_CLASS_BRIDGE from Will
Schmidt
ppc64: fix for boot cpu > 31
ppc64: Report cpus > 32 in proc/cpuinfo correctly
ppc64: merge conflicts
compat_wait4 fix
ppc64: Always pass non segment faults on the 0xc region up to
do_page_fault
ppc64: Add some branch prediction
ppc64: remove ioperm
ppc64: threaded coredump support
ppc64: fix copy_from_user leak, from Milton Miller
ppc64: Add warning for unhandled irqs
ppc64: remove sys32.S
ppc64: ppc64 update
ppc64: copy_tofrom_user exception handling fix from Paul
Mackerras
ppc64: kcore support
ppc64: Increase maximum allocation size to 16MB for largepage
support
ppc64: FORCE_MAX_ZONEORDER is actually order + 1, from Dave
Gibson
ppc64: fix compile error introduced in threaded coredump
patch
fix matroxfb compile on ppc64]
ppc64: Call setup_kcore later in boot, we call kmalloc from
it
ppc64: cputable support from Will Schmidt
ppc64: use an initcall to register ras irqs
update ppc64 MAINTAINERS entry
ppc64: forgot to add this in the cputable merge
ppc64: Fix overallocation of NUMA bootmem bitmap and fix for
NUMA kernels on non NUMA boxes
ppc64: Fix for multiple zone 0 regions on many ppc64 NUMA
boxes
ppc64: Fix ppc64 build
ppc64: Update BUG handling based on ppc32
ppc64: use common 32bit ioctl code
ppc64: K&R to ANSI style conversions from Steven Cole
ppc64: copy_tofrom_user fix from Paul Mackerras
ppc64: various fixes to sys32_sysinfo, from Will Schmidt
ppc64: Rework inline syscall macros, fix clobbers & gcc3.3
from Franz Sirl
ppc64: rework user access functions
Arnaldo Carvalho de Melo:
wanrouter: fix bug introduced by latest namespace fix
net: abstract access to struct sock ->flags
USB drivers: initialize struct usb_driver ->owner field
n_hdlc: CodingStyle cleanups and removal of old stuff
net: reduce the data dependency of struct sock/struct
tcp_tw_bucket
net: create struct sock_common and use in struct sock &
tcp_tw_bucket
list.h: improve hlist
udpv6: use the right struct sock when testing if it is PF_INET6
family
Art Haas:
C99 patches for fs/
C99 struct initializers for kernel files
Bart De Schuymer:
[BRIDGE]: Remove unnecessary code in br_input
[NETFILTER]: Fix ARPT_INV_MASK in arp_tables.h
Bartlomiej Zolnierkiewicz:
allow "hdX=scsi" for modular scsi/ide-scsi
kill "hdX=noremap"
fix two IDE list_head problems
create /proc/ide/hdX/capacity only once
kill recreate_proc_ide_device()
switch ide to taskfile IO
Ben Collins:
Update IEEE1394 (r939)
USB Multi-input quirk
Update IEEE1394 (r946)
[SPARC64]: Final image strip not to strip too much
USB: fix keyboard leds
Register scsi devices after naming them
Update IEEE1394 (r952)
Benjamin Herrenschmidt:
[SUNGEM]: Fix gcc3.3 warnings
Nuke check_highmem_ptes()
Bruce D. Elliott:
[SPARC64]: Fix transmit handling in sunsab.c serial driver
Bruce Fields:
gss_marshal and gssvalidate depend on
gss_credget_ctx never returning NULL; but gss_refresh
depends on gsscredget_ctx returning NULL
whenever the cred is not up to date. So, I replaced the single
gss_credget_ctx by a gss_credget_ctx and a
gss_credgetuptodate_ctx.
add handling of the new CTXPROBLEM and CREDPROBLEM RPCSEC_GSS
errors
Add a "protocol: udp/tcp" line so that gssd can use the same
protocol for null calls that was specified in the mount
options.
I believe we need to set a timeout before doing the sleep in
gss-upcall
allow gssd to communicate failure to initialize contexts back
to the kernel, so the kernel can return -EACCES when a user lacks
credentials, instead of just hanging until they kinit.
This makes several changes to the gss upcalls
Trivial; I kept forgetting what each of the xdr_netobj's passed
to the gss-api routine meant, so I thought I'd fool with the
argument names in an effort to make them more helpful.
Trivial fix for a typo in fs/nfs/nfs4state.c
Chas Williams:
[ATM]: lane and mpoa module refcounting and locking
cleanup
[ATM]: HE driver coding style conformance
[ATM]: HE driver misc irq handler cleanups
[ATM]: Move rategrid off stack in HE driver
[ATM]: more cleanup in HE driver
Chris Wright:
lsm: Early init for security modules (1/4)
lsm: Remove task_kmod_set_label hook (2/4)
lsm: Remove inode_permission_lite hook (3/4)
lsm: setfsuid/setgsuid bug fix (4/4)
Christoph Hellwig:
acpi serial stuff
kill register_pccard_driver
give ->proc_info a struct Scsi_Host * parameter
PPC32: Syscall cleanups
[NET]: Switch lanmedia driver to initcalls
[NET]: Remove sdla from setup.c
[SPARC64]: Kill sys_aplib
use second arg to scsi_add_host in usb storage
fix scsi_register_host abuse in usb scanner drivers
driver model for scsi upper drivers, take 2
LDM-based scsi host lookup
switch /proc/scsi/scsi to seq_file and proper device model
kill scsi_host_get_next
kill remaining direct uses of scsi_register_host
fix signal.h compilation on PPC
[NET]: Move dmascc away from setup.c
[NET]: Fix non-modular sdla.c build
[NET]: Kill useless/wrong Version line from net/core/dv.c
[NET]: net/core/dst.c typo
[NET]: Fix coding style in net/core/filter.c
[NET]: Fix coding style in net/core/iovec.c
[NET]: Convert ppc early-init drivers to initcalls
[NET]: Fix accidental revert of init_etherdev killing in PPC
net drivers
[NET]: Convert skfp over to initcalls, kill fddi cruft from
Space.c
Daniele Bellucci:
USB: replaced BKL in rio500.c
Dave Engebretsen:
[netdrvr pcnet32] bug fixes
Dave Jones:
AM53C974 request region
[AGPGART] Compilation fix
[AGPGART] Remove useless early agp_init() from i810fb
agp_init() just printk's a banner. This is unnecessary at this
early stage.
[AGPGART] Yet another missed typedef compile fix
[AGPGART] Report fixing of errata, and add missing printk stuff
(n's, KERN_INFO, PFX)
[AGPGART] Kill useless printk in frontend
[AGPGART] Missing printk levels
[CPUFREQ] Make powernow-k7 leap big buildings^Wranges
[netdrvr skge] add ULL modifier to 64-bit constant
[netdrvr] add MAINTAINERS entry for atmel wireless driver
Jeff Wiedemeier:
compile fix for agp_memory struct definitition change
Jens Axboe:
remove buggy BUG_ON in ide-cd
blk layer tag resize
copy the tag_map
ide-cd buglets
scsi_ioctl HZ fixes
ide-cd/scsi/block fixups for SG_IO
kill old stuff
Jim Houston:
preallocate signal queue resource - Posix timers
Joe Burks:
USB: vicam.c patch
Joe Thornber:
dm: Replace __HIGH() and __LOW() macros
dm: signed/unsigned audit
dm: new suspend/resume target methods
dm: Lift dm_div_up()
dm: Fix memory leak in dm_register_target()
dm: Remove some debug messages
dm: Remove an old FIXME
John Hawkes:
2.5.70 remove smp_send_reschedule() cruft
John Levon:
OProfile: Export task->tgid in the buffer
OProfile: update Changes
OProfile: remove useless code
OProfile: fix init / exit routine
Jon Grimm:
[SCTP] Verify contents of SACK against underrun
[SCTP] Use slab cache for sctp_chunk &
sctp_bind_bucket
[SCTP] Bug fix for bind_bucket leak & heartbeat error
count
[SCTP] Use non-prefetch list walker for short list
[SCTP] CANT_STR_ASSOC could/should return error up to user.
(ardelle.fan)
[SCTP] Multiple causes may be embedded in an ERROR chunk
(ardelle.fan)
[SCTP] Don't use path thresholds determine the overall error
thresh
[SCTP] Let T5 timer affect SHUTDOWN-PENDING associations
[SCTP] sctp_jitter() needs to protect against div by 0
[SCTP] Various bakeoff fixes
[SCTP] Fix double free of chunk along unexepected INIT
path
[SCTP] Incorrect WORD_ROUND on a network endian value
[SCTP] Special case the handling of HOST_NAME_ADDRESS parm
[SCTP] Hand merge bk pull conflict. sk->state now
sk->sk_state
Justin T. Gibbs:
Aic6xxx and Aic79xx Driver Update
Bump aic79xx driver version to 1.3.9
Jörn Engel:
zlib cleanup: remove FAR macro
zlib cleanup: __32BIT__ and STDC
zlib cleanup: ZEXTERN
zlib cleanup: ZEXPORT
zlib cleanup: z_off_t
zlib cleanup: OF
zlib cleanup: C++ workarounds
zlib merge: turboc
zlib merge: inffast.c
Mark Compaq MAINTAINERS entries stale
zlib cleanup: local
zlib cleanup: Z_NULL removal
zlib cleanup: unnecessary cast removal
zlib merge: return code
zlib merge: avoid 8-bit window errors
zlib changes: memlevel
Fix typo in comment
adjust ppp to zlib change
Krishna Kumar:
[NET]: Initialize sysctl_table to NULL in
neigh_parms_alloc
Kurt Robideau:
Rocketport driver against 2.5.70-bk7
Rocket patch against 2.5.70-bk18
Linus Torvalds:
Remove a few zero-sized files, as noted by David Gibson
Make zlib_inflate look more like ANSI C code
zlib cleanup: final fixups
Remove half-deleted zero-sized sound file
Quick response to de-listing the Compaq FC/RAID controllers
from the MAINTAINERS list. How they're HP, and maintained by
Stephen Cameron.
Re-introduce debugging code in list handling, poisoning stale
list pointers to give us a nice oops if somebody is doing something
bad.
Fix __d_drop() to properly initialize the d_hash fields, so
that __d_drop() can safely be done multiple times on a dentry
without corrupting other hash entries.
Don't make the source checker default path be quite so
hackish.
Avoid warning by using an inline function rather than a macro
for the default "pci_domain_nr()" definition. The inline function
will evaluate the argument.
Make chr_dev_init() happen after PCI init, but before low-level
driver initializations
Fix rcu list poisoning - since another CPU may be traversing
the list as we delete the entry, we can only poison the back
pointer, not the traversal pointer (rcu traversal only ever walks
forward).
Make d_move() be able to gracefully handle the case of the
dentry already being unhashed on entry.
Remove strange and broken ACPI rule from serial Makefile
Linux 2.5.71
Manfred Spraul:
Prepare for page unmapper
Margit Schubert-While:
I2C: add LM85 driver
Mark Haverkamp:
megaraid driver fix for 2.5.70
Mark M. Hoffman:
I2C: fix oops w83781d during rmmod
I2C: more w83781d fixes
Martin Schlemmer:
ethertap: fix struct sock cleanup leftover
Matt Domsch:
dynids: use list_add_tail
dynids: free dynids on driver unload
Matt Porter:
PPC32: Fix a compile error on 4xx embedded PowerPC
PPC32: Resolve colliding includes in boot wrappers
Matthew Dharm:
USB: storage: abort and disconnect handling
USB: storage: collapse one-use functions
USB: usb-storage: fix typo
USB: usb-storage: timeouts and aborts
USB: usb-storage: usb_stor_control_msg() and stuff
USB: usb-storage: change result codes
USB: usb-storage: handle babble
USB: usb-storage: re-organize probe/disconnect
USB: usb-storage: remove dead code
Matthew Wilcox:
PCI: domain support for sysfs
Mike Anderson:
Call release on scsi legacy LLDD
Miles Bader:
Remove some unneeded register saving on the v850
Include <linux/fs.h> in
arch/v850/kernel/rte_cb_leds.c
Miscellaneous v850 whitespace and comment cleanups
Handle new do_fork return value on v850
Add __KERNEL__ guard to nb85e_cache.h on v850
Add leading underline to new linker-script symbols on the
v850
Whitespace and comment cleanups for v850 entry.S
Add v850 support for hardware single-step (via ptrace)
Update irq.c on v850 to use irqreturn_t
const-qualify memory arg in v850's __test_bit
Mitsuru Kanda:
[IPV6]: Fix esp6 extension headers handling
Moritz Mühlenhoff:
[CRYPTO]: Default CRYPTO and MD5 to y if IPV6_PRIVACY is
enabled
Neil Brown:
md: Export bio_split_pool for md to use
md: Use new single page bio splitting for raid0 and linear
md: Handle concurrent failure of two drives in raid5
md: Improve test for which raid1 device doesn't need to be
written to
md: Fix simple off-by-one error in md driver
md: Get rid of vmalloc/vfree from raid0
md: Always allow a half-built md array to be stopped
md: Improve raid0 mapping code to simplify and reduce mem
usage
md: Remove dependancy on MD_SB_DISKS from multipath
md: Remove dependancy on MD_SB_DISKS from raid5
md: Remove dependancy on MD_SB_DISKS from raid0
md: Remove MD_SB_DISKS limits from raid1
md: Remove dependance on MD_SB_DISKS in linear personality
md: Replace bdev_partition_name with calls to bdevname
Fix raid5 bug where wrong 'dev' is used
Fix raid1 handling of writing to multiple devices
Fix up freeing of kmalloc structures
Fix bug in /proc/mdstat
md - Zero out some kmalloced space in md driver
Nicolas Pitre:
[ARM PATCH] 1545/1: correct compiler flags for
ARMv5TE/XScale
[ARM PATCH] 1540/2: fixes for gcc-3.3
Olaf Hering:
USB: incorrect ethtool -i driver name
Oliver Neukum:
fix irq handling for DC395
improve Documentation for DC395
01-debug-cleanup.patch
major cleanup of the module code
USB: allocate memory for reset earlier
USB: return errors when disabling a port
USB: cut usb_set_config from hpusbscsi
USB: usb_set_configuration in empeg.c
USB: kill a compiler warning in hpusbscsi
Patrick McHardy:
[PPP] fix memory leak in ioctl error path
Patrick Mochel:
drver model: Add release method for class devices
sysfs: Fix binary file handling
[driver model] Clean up CPU unregistration
Driver Class: don't call put_device() when we never called
get_device()
[driver model] Clean up class release handling
[kobject] Update Documentation and licenses
[driver model] Update copyrights and license statements
[driver model] fix comment in device.h
[fs] Remove kobject support for filesystems
[kobject] Remove kobj_lock and use lockless refcounting
[driver model] Add device_for_each_child iterator
[kobject] Add warning + back trace if kobject_get() is called
with 0 refcount
[driver model] Remove extraneous class device release
method
[driver model] Rewrite system device API
[list.h] Add list_foreachentry_reverse
[kobject] Add set_kset_name
[driver model] Make sure that system devices are handled
specially power-wise
[driver model] Convert to new system device API
[lapic] Convert to new system device API
[i8259] Convert to use new system device API
[nmi] Convert to use new system device API
[timer] Convert to use new system device API
[oprofile] Convert to use new system device API
[x86-64 i8259] Convert to use new system device API
[s390 xpram] Convert to use new system device API
[driver model] Create include/linux/sysdev.h and define
sysdev_attribute
[driver model] Make sure right header is used for cpu.c
[memblk] Convert to use new system device API
[numa nodes] Convert to use new system device API
[driver model] Remove system device definitions from
device.h
[apic] Use sysdev.h instead of device.h
[i8259] Use sysdev.h instead of device.h
[nmi] Use sysdev.h instead of device.h
[timer] Use sysdev.h instead of device.h
[oprofile] Use sysdev.h instead of device.h
[x86-64 i8259] Use sysdev.h instead of device.h
[s390 xpram] Use sysdev.h instead of device.h
[driver model] Add save() and restore() methods for system
device drivers
[driver model] Don't Oops when registering global sysdev
drivers
[cpu] Use sysdev.h instead of device.h and export
cpu_sysdev_class
[mtrr] Add save()/restore() methods
[cpufreq] Convert to use new system device API
[sysfs] Get zeroed page for file read/write buffers
[netdrvr bonding] fix long failover in 802.3ad mode
[netdrvr bonding] fix ABI version control problem
Simon Kelley:
[netdrvr] add atmel[_cs], new wireless driver
Sridhar Samudrala:
[SCTP] Support for socket options that pass both addr and
associd
[SCTP] Support for SCTP_GETPEERADDR_INFO socket option
[SCTP] Rename struct sctp_protocol as struct sctp_globals and
define macros for all the global fields in the structure.
[SCTP] SCTP_SHUTDOWN_EVENT notification support
[SCTP] /proc interface to display associations/endpoints
[IPV6]: Allow ipv6 fragmentation via ip6_xmit() when ipfragok
is set
Stephen Hemminger:
[BRIDGE]: Make delete bridge work with current unregister
semantics
[NET]: Sysfs netdev cleanup and bugfix
[NET]: Kill deprecated if_port_text and users
[NET]: Fix device unregister in TUN driver
typo in new class_device_release
sb1000 driver bugs
[DECNET]: Fix build warnings
[NET]: Fix sysfs kobj parent refcount handling
[NET]: Cleanup net-sysfs show and change functions
[NET]: Expose alloc_netdev() for use by drivers
[BRIDGE]: Bridge using alloc_netdev
[VLAN]: vlan network device using alloc_netdev
[TUN]: tun using alloc_netdev
[ACENIC]: Convert to alloc_etherdev
[NET]: Dynamic allocation for dummy net device
[IPV6]: Dynamic allocation for SIT net device
[IPV4]: Dynamic allocation for IPIP net device
[IPV4]: Dynamic allocation for IPGRE net device
[NET]: More reasonable error handling in SLIP driver
unload
Steve French:
Make return code on failed cifs mounts more specific and fix
incorrect smb to posix return code conversions
adjust for change of devname to const char (new mount
format)
Steven Cole:
Use '#ifdef' to test for CONFIG options
K&R to ANSI C conversions for zlib
Yet more K&R to ANSI C conversions
More ANSI C cleanup of zlib
[SPARC]: Fix non-ansi parameter lists
Another final K&R to ANSI C cleanup of zlib
Two more sources of "non-ANSI parameter list" warnings
K&R to ANSI conversions for fs/jfs/jfs_dmap.c and
jfs_xtree.c
Steven Whitehouse:
[AX25]: Sanitize ax25 netdevice private handling
Thomas Osterried:
[AX25]: AX.25 bug fixes
Thomas Schlichter:
[NET]: One missed non-netdev SET_MODULE_OWNER case
[NET]: IPSEC protocol module owner cleanup
Thomas Wahrenbruch:
USB: kobil_sct.c added support for KAAN SIM Reader
Trond Myklebust:
Fix udp_data_ready() to use the correct skbuff interface for
extracting the XID. Following the introduction of zero-copy under
UDP, the data may be entirely located in pages under
skb_shinfo(skb)->frags[].
